Behaviour Recognition Using Chronicles

Chronicles are used to describe behaviours, e.g. when (un) desired behaviours should be detected within an event flow, for instance in aerospace system studies. In this paper we first introduce a language to describe chronicles and event related notions, before defining a denotational semantics of a chronicle recognition within an event flow. We present an operational semantics of chronicle recognitions using coloured Petri nets. This modelling is designed to be modular and composible, and to retrieve information on which events in the flow contributed to the chronicle recognition. We prove on some constructs of the chronicle language that our modelling of the recognition using coloured Petri nets yields the expected recognition.